Wn

AEUa

b

c

d

e

a

ARRAA

KLPTPWO

1

co(ihemcn

0d

Agricultural Water Management 98 (2010) 124–133

Contents lists available at ScienceDirect

Agricultural Water Management

journa l homepage: www.e lsev ier .com/ locate /agwat

hole-tree water balance and indicators for short-term drought stress inon-bearing ‘Barnea’ olives

lon Ben-Gala,∗, Dilia Koolb, Nurit Agama, Gerardo E. van Halsemab, Uri Yermiyahua, Ariel Yafea,ugene Presnova, Ran Erela, Ahmed Majdopa, Isaac Zipori c, Eran Segala, Simon Rügerd,lrich Zimmermannd, Yafit Cohene, Victor Alchanatise, Arnon Dagc

Soil, Water and Environmental Sciences, Agricultural Research Organization, Gilat Research Center, Mobile post Negev 2, 85280, IsraelIrrigation and Water Engineering, Wageningen University, The NetherlandsFruit Tree Sciences, Agricultural Research Organization, Gilat Research Center, IsraelLehrstuhl für Biotechnologie, Biozentrum, Universität Würzburg, GermanyAgricultural Engineering, Agricultural Research Organization, Volcani Center, Israel

r t i c l e i n f o

rticle history:eceived 20 April 2010eceived in revised form 10 August 2010ccepted 14 August 2010vailable online 15 September 2010

eywords:ysimeterlant growthranspirationhysiological monitoringater status

lea europaea

a b s t r a c t

Drainage-weighing lysimeters allowed monitoring of water balance components of non-bearing olive(Olea europaea cv Barnea) trees over a 3-month period including short-term events of controlled butsevere water stress. The objective of the study was to evaluate a variety of soil and plant-based waterstatus and drought stress monitoring methods on the basis of tree-scale evapotranspiration (ET). Asthe trees entered into and recovered from water stress, meteorological data, actual ET (ETa), soil watercontent and changes in leaf turgor pressure were continuously monitored. Additionally, midday mea-surements of stem water potential, stomatal conductance, canopy temperature, and quantum yield ofPSII photochemistry were conducted. Diurnal (dawn to dusk) measurements of all the above were madehourly on days of maximum stress. Shoot elongation rate was measured for periods of stress and recovery.Quantum yield of PSII photochemistry, stomatal conductance, and stem water potential all successfullyindicated reductions in whole-tree water consumption beginning at moderate stress levels. These mea-

sured parameters fully recovered to the levels of non-stressed trees soon after water application wasrenewed. Shoot elongation was reduced 25–30% for the 10-day period during and following drought andrecovered thereafter to levels of non-stressed trees. Whole-tree ETa was reduced by as much as 20% evenfollowing full recovery of the leaf level parameters, suggesting reduced canopy size and growth due to thestress period. Non-destructive, continuous (turgor pressure) and remotely sensed (canopy temperature)methods showed promising potential for monitoring effects of water stress, in spite of technological and

nges

data interpretation challe

. Introduction

Irrigation increases yields by allowing greater fruit bearingapacity per tree and per land area and by allowing expansionf olive cultivation into regions too dry for rain-fed productionFernández and Moreno, 1999; Ramos and Santos, 2009). Whilencreasing fruit and oil yields, increased water application oftenas negative effects on olive oil quality parameters (Berenguer

t al., 2006; Dag et al., 2008). Research trends indicate thataximization of yields while insuring high quality oil requires

onditions of moderate water stress, at least during specific phe-ological stages (Ben-Gal et al., 2009). Optimization of irrigation

∗ Corresponding author. Tel.: +972 8 9928644; fax: +972 8 9926485.E-mail address: bengal@volcani.agri.gov.il (A. Ben-Gal).

378-3774/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.agwat.2010.08.008

requiring further attention.© 2010 Elsevier B.V. All rights reserved.

is therefore expected to improve olive cultivation for oil produc-tion, while encouraging efficient use of scarce water resources.Methods for monitoring or recognizing plant water status and/orwater stress conditions represent potentially powerful and impor-tant management tools for both mature, bearing olives, wheremanagement-induced water stress is expected to optimize oilyields and quality, and for olive trees in non-bearing growth stagesfor which maximum vegetative growth is of interest to growers.

Water status can be measured along the soil–plant–atmospherecontinuum and can be quantified through direct measurement ofsoil and/or plant water status as well as by measuring indirect

effects of the stress on plant parameters. Soil-based water sta-tus methods include sampling and gravimetric calculation of soilwater, measures of soil water potential (tensiometers) and variousprobe-based methods for estimating volumetric soil water content(Jones, 2007).

ater M

saatpaStt2ootiTiooatsawc

cdBas1ptacgbw

fpciwtw

odm

A. Ben-Gal et al. / Agricultural W

Plant-based water status monitoring can involve direct mea-urements of physiological responses to water or can insteadttempt to record changes in secondary processes indirectlyffected by (usually less than optimal) water status. Plant responseo water stress can be measured both at small scales (stem, leaf,arts of leaves), where physiological processes are affected andt the whole-tree-scale, where accumulative effects are evident.mall-scale monitoring includes the measuring of such parame-ers as xylem (sap) flow, stomatal conductance, photosynthesis, leafemperature, and stem, leaf or fruit shrinkage (Jones, 2007; Naor,006). These provide an assessment of the physiological severityf water stress and contribute to the determination of thresh-ld levels, as plants enter into, physiologically “shut down” dueo, and recover from water stress. Physiological-based monitor-ng, however, does not permit quantification of cumulative effects.hese include less than optimum growth, changes in phenolog-cal processes, and subsequent altered plant size and vegetativer reproductive production and must be considered at tree andrchard scales. Tree level quantification of water stress in bothbsolute and relative terms and its effects on growth, produc-ion and water consumption is vital for determining managementtrategies. Tree-scale information is also necessary for developingnd utilizing transpiration driven growth models (De Wit, 1958)here reductions in actual evapotranspiration (ETa) are related to

orresponding reductions in biomass, canopy and yield.Assessment of whole-tree responses to water stress, including

umulative effects on biomass and yield, can be aided by continuousetermination of water balance components using lysimeters (Vanavel, 1961; Hillel et al., 1969; Ben-Gal and Shani, 2002; Marek etl., 2006). Despite several disadvantages inherent to lysimeters, e.g.,idewall-boundary effects and microclimatic effects (Bergstrom,990; Flury et al., 1999; Corwin, 2000), they uniquely allow com-arison of actual water and solute balance between differentreatments under well-defined conditions. Weighing lysimetersdditionally enable high temporal resolution (hours, days) quantifi-ation of ET rates. The information concerning water consumption,rowth and production generated from lysimeter studies cane used to identify and quantify desirable management-inducedater stress levels.

Since such whole-tree-scale monitoring is not practicallyeasible in commercial settings, the assessment of accessible,hysiological-based measurements as indicators of water statusan be valuable for managers and decision makers. The understand-ng of the physiological scale processes and responses in terms of

hole-tree-scale effects of water stress will hopefully lead to prac-ical and feasible field scale methods for management of desired

ater stress levels in olive trees.

The absolute values of measured water status are a function bothf available water for uptake and transpiration, and of climatic con-itions. In order to evaluate plant water status, and especially to useonitored water status data for irrigation scheduling, some type

Fig. 1. Olive lysimeters at Gilat Research Center, Israel. Overview (left)

anagement 98 (2010) 124–133 125

of normalization, either to non-stressed conditions or, preferably,to specific climate conditions, is necessary (Fernandez and Cuevas,2010; Jones, 2004; Jones, 2007; Naor, 2006).

We have used drainage-weighing lysimeters to measure waterbalance components of non-bearing olives (Olea europaea cvBarnea) over a 3-month period including short-term events of con-trolled but severe water stress. The objective of the study was touse calculated actual tree-scale ET as a basis for evaluation of avariety of soil and plant-based water status and drought stressmonitoring methods. Both conventional methods, such as stom-atal conductance and stem water potential, and two novel methodsunder development, measuring leaf turgor pressure and canopytemperature, were evaluated.

2. Materials and methods

Single 2-year old ‘Barnea’ olive tree were planted in fif-teen 2.5 m3 volume free-standing lysimeters (Fig. 1) at the GilatResearch Center in the northwestern Negev, Israel (31◦20′N, 34◦40′

E) in June 2008. The trees were irrigated daily, with quantitiesalways exceeding (by ∼20%) the previous day’s transpiration ratesas calculated from the weight data of the lysimeters. Nitrogen,phosphorus, potassium and micronutrients were added to the irri-gation solution as liquid commercial 7:3:7 (N:P2O5:K2O) fertilizerat a rate of 50 ppm N. Each lysimeter (Fig. 1) consisted of a polyethy-lene container (1.4 m high × 1.5 m diameter) filled with loamy sandsoil, a bottom layer of highly conductive porous media (rock-wool) in contact with the soil, and drainage piping filled with thesame material extending downward from the lysimeter bottom.The rockwool drainage extension (Ben-Gal and Shani, 2002) dis-allowed saturation at the lower soil boundary while permittingwater to move out of the soil and be collected. The lysimeter systemincluded automatic water and fertilizer preparation and deliveryand automatic drainage collection as described by Tripler et al.(2007). The soil surface in the lysimeters was covered by a waterpermeable geotextile (Non-Woven Geotextile, 500 g m−2, Noam-Urim, Israel) to minimize evaporation losses. Individual lysimeterswere positioned on square weighing platforms with load cells sit-uated in each corner. By distributing load cell output current onlyover the relevant range of interest (4–5 tons) a resulting resolu-tion of ±15.5 g was reached. ET was calculated daily accordingto: ET = I − D −�W; where I is irrigation, D is drainage and �W ischange in soil water. There was no rainfall during the experimentalperiod.

The lysimeters were divided into three groups (Group 1, Group2 and Group 3) of five trees. All trees received identical treatment

from planting until the beginning of the current experiment. Tostudy the effects of short-term stress, irrigation to each group oftrees was withheld for a number of days during the summer of2009. A preliminary stress round was conducted from the 4th tothe 13th of June on Group 1. Based on the results from this stress

and individual single tree weighing-drainage lysimeters (right).

1 ater M

psGtmtac(tewattmd(ucDteoPGdioomdmd(i

0seaeafec

TStd(t

26 A. Ben-Gal et al. / Agricultural W

eriod, along with estimations of anticipated changes in canopyize and climate demand, irrigation was withheld for 6 days forroups 2 and 3, from July 23rd to July 28th, and from August 28th

o September 2nd, respectively. Continuous monitoring included:eteorological data acquired from a weather station 200 m from

he site, actual ET (ETa) computed from the lysimeter water bal-nce, and relative changes in leaf turgor pressure with leaf patchlamp pressure probes (Zimmermann et al., 2008). Reference ETET0) was calculated according to the FAO Penman-Monteith equa-ion for hourly time steps as described by Allen et al. (1998). Inach period of drought and recovery, additional water status andater stress measurements were conducted daily between 12:00

nd 14:00 local time on all trees. The “recovery” period referso the time it took for all physiological parameters to return tohe reference values or to stabilize on new values. The additional

easurements included volumetric soil water content from timeomain reflectometry (TDR) using a Tektronix 1502 cable testerTektronix Inc., Beaverton, OR, USA), stem water potential ( stem)sing a Scholander type pressure chamber (MRC, Israel), stomatalonductance (gS) from a diffusion leaf porometer (SC-1, Decagonevices, Inc., Pullman, WA, USA), canopy temperature (TC) from

hermal images acquired by an un-cooled infrared thermal cam-ra (ThermaCAM model SC2000, FLIR systems), and quantum yieldf photosystem II photochemistry (˚PSII) using a portable MINI-AM photosynthesis yield analyzer (Heinz Walz GmbH, Effeltrich,ermany). Vegetative growth (shoot elongation) was measureduring and after the stress period of Group 3. The daily monitor-

ng continued after re-watering until full recovery was observed. Inrder to capture the diurnal behavior of the different parametersn both well-irrigated and drought-stressed trees, the measure-ents were augmented and conducted throughout the day on the

ay before re-watering (day of minimum available water, mini-um ET and maximum water stress). Diurnal monitoring on these

ays started from before dawn (4:00) and continued until after dusk20:00). A summary of the experimental setup and dates is givenn Table 1.

Tree-based measurements were conducted on stems and leaves.5–1.5 m above the soil surface. Stem water potential was mea-ured as described by Shackel et al. (1997) on single shootndings with 6–7 leaves covered at least 2 h in advance by sealedluminum–plastic bags. Shoot endings were taken from the north-rn (shaded) side of the trees’ canopies. Stomatal conductance

nd fluorescence-based measurements were taken on young butully grown leaves between 5 and 20 cm from the shoot tip. Forach tree five samples, uniformly distributed over sun-exposedanopy, were taken. Shoot growth was measured by marking a

able 1ummary of drought periods and measurements taken. Whole-tree evapotranspira-ion rate (ETa), stem water potential ( stem), stomatal conductance (gS), area underaily pressure probe data (PPA), quantum yield of photosystem II photochemistry˚PSII), canopy temperature (TC), crop water stress index (CWSI), and shoot elonga-ion (shoot length).

Group under stress 1 2 3

Designation Preliminary July AugustDate irrigation discontinued 04 June 23 July 28 AugustDate irrigation renewed 14 June 29 July 03 SeptemberDate of diurnal measurements − 28 July 02 September

Measurements takenETa + + + stem + + +gS + + +PPA + + +˚PSII − − +Tc and CWSI − − +Shoot length − − +

anagement 98 (2010) 124–133

point 20 cm from the tip of ten shoots equally distributed aroundeach tree on the day that irrigation was discontinued for Group 3.Increase in shoot length was recorded on the 10th and 20th dayfollowing.

The efficiency of photosystem II photochemistry was deter-mined by computing ˚PSII according to the relationship:˚PSII = (F′

m − Ft)/F′m, where F′

m is maximum fluorescence whenapplying a saturation flash to a light-adapted leaf; and Ft is thelevel of fluorescence immediately before saturation flash (Maxwelland Johnson, 2000).

Leaf patch clamp pressure probes (LPCP probes) were used tomonitor relative changes in leaf turgor pressure. Three trees ineach group were each equipped with single LPCP probes, locatedinside the canopy to avoid direct sunlight. The principle of the probeis fully explained in Zimmermann et al. (2008) and Westhoff etal. (2009). In short, two magnets were used to apply a constantknown pressure to a young fully developed leaf and to sense thepressure passing through the leaf. The magnitude of the sensedpressure depends on a turgor-independent, constant term (aris-ing from the compression of the cuticle, air-filled spaces, cell wallsand the silicone of the probe) and turgor pressure. Thus, if theexternal pressure is kept constant and an equilibrium value hasbeen reached after clamping, the difference between the exertedand the sensed pressure only depends reciprocally on turgor pres-sure (Zimmermann et al., 2008). Probe output pressure valuesdecrease to minimum values at nighttime when turgor is high-est and increase during the day with the diminishing of turgorin leaves. Under less than optimum water conditions, both max-imum and minimum output pressure increase, indicating lowerdaytime turgor pressure and less nighttime recovery, respectively.An example of a time course for a single probe before, during,and after a drought event is shown in Fig. 2. We used the cal-culated area under daily time course curves of probe pressurereadings (PPA), as highlighted in the figure, to give a relative quan-tification of turgor-related water stress. PPA data during droughtand recovery periods were normalized to non-stressed valuesand taken as 100% on the day immediately prior to droughtinitiation.

Thermal images were taken 2 m above the canopy of individualtrees. Each image contained a wet reference to aid in normaliz-ing temperature to climatic conditions. The images were used todetermine both absolute average canopy temperature and climate-normalized crop water stress index (CWSI; Jones, 1992) for eachtree, according to the methodology described by Ben-Gal et al.(2009).

Statistical variance, regression and correlation analysis wasaided by JMP (SAS Institute Inc., Cary, NC, USA), SPSS 15.0.1 forWindows (SPSS Inc., Chicago, IL, USA) and SigmaPlot 10.0 (SystatSoftware, San Jose, CA, USA).

Fig. 2. Raw data from a single leaf clamp pressure probe on a tree in Group 2. Periodof ceased irrigation is enclosed by a block. Examples of area under the graph forsingle days (PPA) are shadowed for a day with full irrigation (22 July) and on thethird day without irrigation (25 July). The arrow represents reaching conditions ofapparent near-zero turgor and onset of inversed diurnal probe output data.

ater M

3

3

bAaEeetadpAig3rs1

FGe(

A. Ben-Gal et al. / Agricultural W

. Results

.1. Evapotranspiration and growth

Reference ET (ET0, Fig. 3A) averaged ∼10 ± 1 mm day−1 from theeginning of the experimental period until mid-August. From mid-ugust till the end of the experimental period, ET0 declined linearlynd reached ∼7 mm day−1 at the end of September. Whole-treeT (ETa) was calculated daily using water balance over the entirexperimental period, starting in June and continuing through thend of September 2009. Fig. 3B depicts average daily ETa for each ofhe three groups. Evapotranspiration increased over time with dipspparent during the stress periods. Also visible in Fig. 3B are a 3-ay period from 30 June–1 July with particularly high ETa reflectingarticularly high ET0 (Fig. 3A) and a slight dip at the beginning ofugust when problems with the irrigation system caused delayed

rrigation events and several days of reductions in ETa of all theroups. During their respective drought periods, Group 2 and Group

exhibited nearly identical patterns of reduction in ETa. After a

ecovery period of 4 days, a permanent decrease in the ET of thetressed trees was evident compared to non-stressed trees. Groupexperienced relatively slight stress in June’s preliminary drought

ig. 3. Evapotranspiration (ET) over the time course of summer 2009 including droughtroups 1, 2 and 3, left to right, respectively. Symbols are averages and error bars standarvapotranspiration (ETa), (C) accumulated average ETa (measurements began from timm2/tree) calculated by normalizing actual ET with reference ET (ETa/ET0).

anagement 98 (2010) 124–133 127

period which caused a visible, but non-significant retardation in ETa

at the early stage of the experiment. The ETa of Group 1 was not per-manently affected and returned to that of non-stressed trees duringthe subsequent months. Daily values of ETa for Group 1 nearlytripled over the course of the summer, from just over 30 L tree−1

to more than 80 L tree−1. Maximum values of daily ETa for Groups2 and 3, which experienced more severe water stress during theirrespective drought periods, were significantly lower than those ofGroup 1 and reached approximately 70 L tree−1. Evapotranspira-tion for all the groups stopped increasing towards the end of Augustand began to decline after the final stress period in mid-September.

Actual ET as seen in Fig. 3B was a function of canopy size andenvironmental evaporative demand in addition to water avail-ability. In order to aid interpretation of the data and to facilitatesubsequent discussion, the ETa data have been manipulated in twodifferent ways in Fig. 3C and D. First, ETa was accumulated over thecourse of the experimental period. Total average accumulated ETa

per tree (Fig. 3C) was around 3200 L for all groups at the beginning

of the summer. This number is the sum of daily tree-scale ETa start-ing from the planting of the trees in June 2008. Deviations of thepatterns of accumulated ETa over the course of the summer of 2009(Fig. 3C) were evident during and following the drought periods.

periods. Periods of discontinued irrigation are shown as gray horizontal bars ford deviation, n = 5. (A) reference evapotranspiration (ET0), (B) daily actual tree-scalee of planting in lysimeters: July 2008), and (D) effective transpiring canopy area

1 ater M

Dacio7li1it(ilAasmiuassn

a

Fs(gr3lfn

28 A. Ben-Gal et al. / Agricultural W

uring drought periods the decreased ETa resulted in lower totalccumulated ETa, and decreases were also found in the slopes of theurves in Fig. 3C following each group’s period of drought. Follow-ng the drought period to Group 2 (11 July–01 September) the slopef its accumulated ETa curve was 66.1 L tree−1 day−1 compared to2.3 and 72.1 L tree−1 day−1 for Groups 1 and 3 respectively. Fol-

owing the drought period for Group 3 (14 September–01 October)ts slope was 55.5 L tree−1 day−1 while that of the control (Group) was 69.9 L tree−1 day−1. The second manipulation was comput-

ng the average effective transpiring canopy size of the individualrees (m2 tree−1), calculated by dividing ETa (L tree−1 day−1) by ET0mm day−1 = L m−2 day−1) (Fig. 3D). Changes in effective transpir-ng canopy size reflect tree growth as long as transpiration is notimited by water availability or some other stress-causing factor.pparent reductions during periods where irrigation was stoppedre obvious, as is a slight dip at the beginning of August during thehort period of reductions in ETa of all the groups due to systemalfunction and delayed irrigation events. The effective transpir-

ng canopy area increased for all groups throughout the periodp until mid-September, when it more or less stabilized. Startingt around 4 m2 tree−1, effective transpiring canopy areas of non-

tressed trees reached more than 11 m2 tree−1 by the end of theummer while the trees of Groups 2 and 3, experiencing the pro-ounced drought periods, reached a maximum of 9 m2 tree−1.

Fig. 4 provides a look at the dynamics of climate (ET0), soil water,nd ETa during the drought and re-irrigation periods. Diurnal pat-

ig. 4. Hourly reference evapotranspiration (ET0) rates for both stress periods (A),oil water from TDR measured at 4 depths (20, 40, 60 and 80 cm from soil surface)B), measured during 26 August–06 September; and evapotranspiration (ET) (C),oing into and coming out of stress periods. The X-axis is days of stress (S) andecovery (R) for Group 2 (23 July–03 August; indicated as “July”) and for Group(26 August–06 September; indicated as “August”). WW stands for well-watered

ysimeters and D for lysimeters subjected to drought. ET values are average ratesor five trees. TDR values are from individual probes in representative irrigated andon-irrigated lysimeters.

anagement 98 (2010) 124–133

terns of ET0 (in units of mm h−1) are displayed for the two periods inFig. 4A. In the figure, the greater climate demand in the July com-pared to the August period is evident in the slightly longer daylength with positive ET0 and significantly greater peak rates. Soilwater content measured at four depths in two individual lysimetersfrom 26 August till 6 September is shown in Fig. 4B. Volumetric soilwater content in the upper layers of the well-watered lysimeterfluctuated from ∼20% immediately following irrigation to less than10% prior to the daily irrigation events. Lower soil layers showedsimilar wetting and drying patterns with smaller fluctuations. Thesoil water content of the lysimeter subjected to drought, whichwas identical to that of the well-watered lysimeter prior to stop-ping daily irrigations, declined to 5–10% after 1 day of drought andto 2–6% on the ultimate day of drought, after 6 days without irri-gation. During the drought period, volumetric water content waslower in the upper soil layers compared to closer to the bottom ofthe lysimeter. Tree-scale hourly ETa rates calculated from lysime-ter water balance are shown in Fig. 4C for both the “July” (23 July–3August) period when Group 2 trees were induced to water stressand for the “August” (26 August–6 September) period when Group3 trees experienced drought. Daily patterns for both well-wateredand drought-stressed trees were remarkably similar during the twoperiods. Rates and patterns of ETa were unaffected on the first dayirrigation was discontinued and reduced steadily thereafter. On thesecond and third days following, the non-irrigated trees followedthe same daily parabolic curves as the non-stressed trees, but withlower peaks. From the fourth day, peak ETa rates shifted to ear-lier in the day. Midday decreases in ETa were found for stressedtrees and for the well-watered trees in July. The midday reduc-tions began earlier and lasted for a longer time compared to thosein well-watered trees during the stress period and the first fewdays after re-irrigation during the recovery period. Rates of ETa

increased to stable values following re-introduction of irrigation,but to levels lower than those of continuously irrigated trees. Weused the data from Fig. 4 to compare water consumption calcu-lated from the soil-based TDR monitoring to the water balanceETa calculations for the non-irrigated trees. Rough calculation ofwater in the root zone from the TDR data (approximate total soilvolume of 2 m3 and decrease of depth averaged volumetric watercontent from 15% to 4%) indicated a loss of around 220 L. Accumu-lated ET loss during the same period (integral of the area underthe ETa curves in Fig. 4B) was 203 L for Group 2 and 232 L forGroup 3.

The apparent decelerated canopy growth rates found in thewhole-tree ETa values (Figs. 3 and 4) during and following droughtperiods were also evident in shoot growth measurements. Shootlength was measured 1 day before irrigation to Group 3 wasdiscontinued (27 August) and elongation was measured after 10days and again after 20 days. Analysis with one way ANOVA andTukey–Kramer all pairs test showed significantly less growth forGroup 3 (2.3 cm) compared to Groups 1 and 2 (3.7 and 3.3 cm,respectively) for the 10-day period including the water stress andrecovery. For the next 10-day period, shoot growth in Group 3(2.2 cm) no longer statistically differed from Groups 1 and 2 (2.8and 2.4 cm, respectively).

3.2. Water status and stress monitoring

Differences between irrigated and non-irrigated trees were visi-ble from the first or second day of stress and all parameters returnedto levels corresponding to those of the well-watered trees values

within 4 days after resuming irrigation (Fig. 5). On the sixth and finalday without irrigation, when the trees were under severe stresswith ETa less than 0.15 of irrigated trees and very little remain-ing available water in the soil (Fig. 4), visible leaf curling becameevident.

A. Ben-Gal et al. / Agricultural Water M

Fig. 5. Midday measurements going into and recovering from drought. Stress periodfor Group 3, August 2009. (A) whole-tree evapotranspiration (ET), (B) stem waterpotential ( stem), (C) stomatal conductance (gS), (D) area under daily values fromleaf clamp pressure probe data (PPA), (E) quantum yield of photosystem II photo-ciS

fpti9vtd

hemistry by fluorometer (˚PSII), (F) canopy temperature (TC), (G) crop water stressndex (CWSI). Dotted vertical lines indicate beginning and end of drought period.ymbols are averages and error bars standard deviations, n = 5.

The patterns of response of ETa, stem, gS, and PPA were similaror the drought periods of both Groups 2 and 3 and we thereforeresent them together while including data graphically from onlyhe second period. Well-watered trees had average daily ETa rang-ng from 70 to 90 L tree−1 during the July period and from 80 to

0 L tree−1 during the August period. Their midday average stem

alues ranged from −1.4 to −2 MPa in the first period and −1.3o −1.4 MPa in the second and their corresponding average mid-ay gS rates were 90–125 mmol m−2 s−1 in the first period and

anagement 98 (2010) 124–133 129

90–150 mmol m−2 s−1 in the second. On the first day of droughtthe differences between irrigated and non-irrigated groups werenot yet significant. On the second and third day the ETa of non-irrigated trees dropped about 10% per day, and decreased to about30% of that of irrigated trees on the fourth day. ETa continued todecrease to a low of 14% on the sixth day of stress, after which irri-gation was resumed (Fig. 5A). Similar to the ET, decreases in stem

sharpened after 2 days (Fig. 5B). stem values continued to decrease,to lows of around −3.5 MPa on the last days of stress. Midday gS ofnon-irrigated trees behaved virtually identically to daily ETa, andreached low points of ∼30 mmol m−2 s−1 on the last days of stress(Fig. 5C). ETa stabilized after 3–4 days following resumption of irri-gation but remained lower than that of the continuously irrigatedgroups. Stem water potential recovered completely within 1 dayfollowing re-irrigation while gS recovered to non-stress values onthe third day of re-irrigation.

The leaf patch clamp pressure probe raw data (Fig. 2) indicatedlower daytime turgor and only partial nighttime recovery to fullturgor on the second day after cessation of irrigation. The areabelow daily curves from the pressure clamp probes (PPA, Fig. 5D)correlated well with ET and the other measurements, increas-ing during the drought period and decreasing during recovery.The diurnal patterns of the pressure clamp probes inversed undersevere stress levels on the fifth and sixth days of stress duringboth periods (see arrow in Fig. 2), rendering PPA values for thosedays meaningless. After reinstating irrigation PPA values returnedto non-stress levels within 1 day, simultaneously with stem.

The fluorescence-based ˚PSII (Fig. 5E) of non-irrigated treesdecreased to significantly lower values than those of irrigated treesafter 4 days. The low point for˚PSII was reached after 5 days and didnot decrease further on day 6.˚PSII recovered simultaneously with stem. Canopy temperature and CWSI (Fig. 5F and G, respectively)also reflected the effect of drought with higher leaf temperaturescausing higher CWSI values from the second day of drought. On thefourth day of recovery the drought-stressed trees returned to TCand CWSI values similar to those of the well-watered trees.

Diurnal measurements for water status monitoring are foundin Fig. 6 (Group 3 on sixth day without irrigation). The measure-ments of stem and gS closely follow those of ETa throughout theday. Evapotranspiration was negligible in the night and, along withET0, followed parabolic curves during the day peaking between12:00 and 14:00. The diurnal patterns for the July period (Group2 on sixth day without irrigation, data not shown) for ETa, stem,gS, and PPA were similar to those shown for the second period inFig. 6 with a number of climate-related differences apparent. Day-light ET0 (area under relevant curve in Fig. 6A) was 2 mm higheron 26 July as compared to 3 September. Corresponding differenceswere found for stem and gS measurements in irrigated trees; stem

decreased to −2 MPa on 26 July, whereas on 3 September it did notfall below −1.5 MPa, and a midday dip in gS was evident on 26 Julyand not on 3 September. The ETa and gS of the stressed trees slightlyincreased above nighttime values in the mornings, reaching peakvalues between 8:00 and 10:00 and decreasing at noon (Fig. 6Aand C). Similarly, stem decreased from maximum nighttime valuesduring the day (Fig. 6B). Significant differences between irrigatedand non-irrigated trees were found throughout the day for stem

and from 7:00 to 7:30 until at least 18:00 for ETa and gS. Abso-lute values of ETa and gS were similar for the non-irrigated treeson the two dates, ranging, respectively, from 0 to 9 L h−1 tree−1 andfrom 25 to 140 mmol L−1 s−1. Nighttime stem of stressed trees waslower (<2.5 MPa) on 26 July compared to 3 September (<2.0 MPa).Daytime stem of the non-irrigated trees fell to around −3.5 MPa

on both days and remained fairly stable between 10:00 and 15:00.˚PSII (Fig. 6D) showed significant differences between irrigated andnon-irrigated trees between 9:00 and 16:00 on 3 September. Sim-ilarly TC and CWSI (Fig. 6E and F) showed significant differences

130 A. Ben-Gal et al. / Agricultural Water Management 98 (2010) 124–133

F sed (G( f phoc tions

baT

Eddc

4

a

ig. 6. Diurnal measurements for well-watered (Groups 1 and 2) and drought-stresB) stem water potential ( stem), (C) stomatal conductance (gS), (D) quantum yield orop water stress index (CWSI). Symbols are averages and error bars standard devia

etween irrigated and non-irrigated trees, with largest differencesround midday and more pronounced differences in CWSI than inC.

Significant correlations were found between whole-tree dailyT rates and midday values of stem, gS, ˚PSII, and PPA for theata sets entering into and recovering from drought (Table 2). Mid-ay canopy temperature and CWSI were less but still significantlyorrelated with daily whole-tree ET.

. Discussion

The behavior of groups of drought-exposed trees during the Julynd August periods of ceased irrigation was coincidentally identi-

roup 3) olive trees on September 2, 2009. (A) whole-tree evapotranspiration (ET),tosystem II photochemistry by fluorometer (˚PSII), (E) canopy temperature (TC), (F), n = 5.

cal. The trees had a given, fixed, root zone volume of approximately2 m3 of soil, dictated by the lysimeter containers. Thus at the startof a drought experiment there was a fixed amount of water in thesoil which could be depleted. The rate of ETa (Figs. 3B and 4C), andtherefore the rate of depletion of soil water (Fig. 4B), was dependentupon climate demand and tree size. In Group 2’s period of drought,the climate demand was higher than in Group 3’s (∼10.0 mm day−1

compared to ∼8.0 mm day−1 ET0) but this was compensated by

the fact that the trees had grown during the month between thedrought periods. Therefore, continuously well-irrigated trees hadeffective transpiring canopy areas (Fig. 3D) that were some 40%larger (10 m2 compared to 7 m2) at the start of the Group 3’s stressperiod compared to the beginning of the Group 2’s. The contingent

A. Ben-Gal et al. / Agricultural Water Management 98 (2010) 124–133 131

Table 2Regression coefficient (r2) values between the water stress indicators for daily measurements going into and recovering from drought: whole-tree evapotranspiration (ETa),stem water potential ( stem), stomatal conductance (gS), quantum yield of photosystem II photochemistry (˚PSII), canopy temperature (TC), crop water stress index (CWSI)and area under daily pressure probe data (PPA).

ETa � stem gS ˚PSII TC CWSI

� stem 0.69 ***

gs 0.72 *** 0.58 ***

˚PSII 0.65 *** 0.63 *** 0.49 ***

TC 0.46 ** 0.5 ** 0.56 ** 0.34 *

CWSI 0.38 * 0.25 ns 0.33 * 0.23 ns 0.55 **

PPA 0.59 *** 0.3 *** 0.26 ** 0.26 * 0.13 ns 0.17 ns

n

csp

TrGesdmcncSf

ptcrttottordaotoawab

ct(imporaFe p

s = not significant.* P ≤ 0.05.

** P ≤ 0.01.*** P ≤ 0.001.

onsequence of this was that absolute values of ETa and the rate ofoil water depletion were virtually identical in both experimentaleriods.

The short-term periods of drought decreased vegetative growth.his was apparently true even for Group 1 that experienced only aelatively mild stress in the preliminary experiment in June. Whileroup 1’s ETa rates after recovery were not significantly differ-nt from the other groups, patterns of accumulated ET impliedmaller trees and smaller growth rate following the preliminaryrought period. The growth rate of the young trees during the sum-er of 2009 was fairly remarkable. Climate-normalized ET-based

alculations (ETaET0−1: Fig. 3D) indicate that non-stressed trees

early tripled their canopy area between June and October. Theanopy growth driven increase in ET stopped towards the middle ofeptember, after which the transpiration stabilized around a fixedraction of climatic demand (ET0, Fig. 3D).

The single episodes where available soil water was nearly com-letely depleted caused long-term significant decreases of morehan 20% in transpiration rates and in the effective transpiringanopy of individual trees. This impairment of whole-tree transpi-ation following the periods of drought was not reflected in any ofhe leaf or canopy-based measurements, all of which recovered tohe values of trees that did not experience stress within a numberf days following re-irrigation. This supports the claim that long-erm reduction of ETa was due simply to reduced growth duringhe drought itself and to the subsequent lower growing capacityf the smaller trees. Similar reductions in vegetative growth wereeported in field experiments by Greven et al. (2009) for a 3-monthrought period on ‘Verdale’ olives in New Zealand and by Iniesta etl. (2009), for 2 years of continuous deficit irrigation of ‘Arbequina’lives in Spain. Shoot growth was also found to be highly sensi-ive to drought by Sofo et al. (2008) who compared dry weightsf shoots in rain-fed and irrigated ‘Coratina’ olives in Spain over7-year period. In that study partial inhibition of shoot growthas reported at pre-dawn leaf water potentials as high as −1 MPa,

nd severe inhibition was found at pre-dawn leaf water potentialselow −1.3 MPa.

All the measured parameters were sufficiently sensitive toonditions causing lower than optimal water consumption andherefore are potentially useful for monitoring water statusTable 2, Fig. 5). The coincidental similarity in developmentnto and recovery from stress experienced by Groups 2 and 3

ade direct comparison of day by day response of physiologicalarameters possible and aided in correlating between the vari-us methods. Positive correlations were found for the data setepresenting the midday values for all the trees entering into

nd recovering from drought stress (Table 2). It is obvious fromig. 5 that the methods are all better correlated for the periodntering into stress compared to the period of recovery wherestem and PPA return to non-stressed levels faster than the otherarameters.

Entrance into stress (reduction of ETa) was identified by signif-icant differences between midday measurements of irrigated andnon-irrigated trees for stem, gS, TC, and CWSI. The fluorescence-based photosynthesis calculations were slightly less sensitive toinduction of water stress as they showed significant differencesa day later. Recovery following the stress was also successfullyportrayed by all the methods, each of which was returned to non-stressed values. stem, PPA and ˚PSII fully recovered after 2 daysof irrigation while gs and the canopy temperature-based measure-ments more closely followed ETa and returned to steady high valuesafter 4 days.

The turgor-related PPA measurements in response to stresswere relatively insensitive to climate and gave the same level ofresponse during the two drought periods. Changes in PPA enter-ing into drought were highly pronounced compared to the otherparameters, substantiating that turgor is particularly sensitive tochanges in water potential of the soil (Hsiao, 1973). This was alsoreported by Sofo et al. (2008) who found onset of partial and severeinhibition of turgor pressure at pre-dawn leaf water potentials of−1 and −3.2 MPa respectively in ‘Coratina’ olives planted in pots.For mild to moderate stress levels in particular, PPA seems anappropriate measure for water status, especially since the methodis non-invasive, continuous and automatic. At near-zero turgor theassumptions leading to a sensor’s constant term were apparentlyno longer valid and the resulting calculated PPA from output valueswas meaningless, but this effect was reversible. Upon re-watering,increased relative turgor pressure was measured as seen by thereturn to the accustomed profiles of the probe output pressures andto pre-stress levels of calculated PPA. Normally, such severe stresslevels and zero leaf turgor pressure values would not be considereddesirable and therefore their measurement is likely not necessaryfor monitoring and water management purposes.

Absolute values of the other physiological-related monitor-ing methods were affected by climate as well as by water statusand were therefore different for the two periods. The values ofmeasured stomatal conductance were particularly influenced byclimate. Measurements from irrigated trees demonstrate that gS inthe first period, with a higher climatic evaporative demand, wasonly around 120 mmol m−2 s−1 compared to 150 mmol m−2 s−1

in the second period. These numbers are lower than thosereported for well-watered trees by Pérez-López et al. (2008)(190 mmol m−2 s−1) and Giorio et al. (1999) (240 mmol m−2 s−1).The gS of severely drought-stressed trees, ∼30 mmol m−2 s−1 dur-ing both stress periods in the current study, were equivalent tothose reported for drought-stressed trees in the previously men-tioned studies. Stem water potential was also affected by climate,

although less pronounced than stomatal conductance. The effect ofclimate on stem decreased as the level of stress increased.

Stomatal conductance was previously found to be correlatedwith stem water potential of olive trees (Ben-Gal et al., 2009;Fernandez et al., 2006; Tognetti et al., 2005, 2007; Pérez-López

1 ater M

e(sw−(ts−l2

rtstnawawt

wactuc(fcti

ssptd

ts2tnscwtws

ctat‘mdG2s2

t

32 A. Ben-Gal et al. / Agricultural W

t al., 2008; Giorio et al., 1999) and of other deciduous fruit treesNaor, 2006). Greven et al. (2009) found low correlation betweentomatal conductance and stem water potential, but were workingith more severely stressed trees with stem water potentials of4 MPa. Stomata are all but closed at −2.5 MPa in many fruit trees

Naor, 2006). We found that transpiration and photosynthesis con-inued for olives, albeit at reduced rates and with diurnal patternshifting to the morning hours, at stem water potential as low as3.5 MPa. Others have suggested that this is true at potentials as

ow as −6 MPa (Sofo et al., 2008) or even −10 MPa (Guerfel et al.,009).

Fluorescence-based measurements are generally understood toepresent particularly sensitive elements of the photosynthesis sys-em (Maxwell and Johnson, 2000). The fact that ˚PSII respondedlower than other parameters as drought was introduced supportshe suggestion that olives have mechanisms to prolong mainte-ance of photosynthesis under stress-causing conditions (Sofo etl., 2008). The measured maximum˚PSII of 0.86 at pre-dawn agreesith the maximum ˚PSII of 0.83 found in most species (Maxwell

nd Johnson, 2000). Quick recovery in photosynthetic efficiencyas also observed by Sofo et al. (2008), who found that photosyn-

hesis recovered even faster than leaf water potential.Canopy temperature was found to be significantly correlated

ith ET, stem, gS, and ˚PSII, and the CWSI to all except for stem

nd ˚PSII. Ben-Gal et al. (2009) showed good correlation betweenanopy temperature and CWSI and soil water, stem water poten-ial, and stomatal conductance in mature ‘Barnea’ olive trees grownnder various levels of sustained deficit irrigation in a commer-ial orchard using images of individual trees. Sepulcre-Cantó et al.2006) found similar correlations for stem water potential and TCrom airborne images of entire orchards. The somewhat weakerorrelations of the CWSI (compared to correlations to TC) imply thathe normalization procedure, originally developed for field crops,s not as effective for olive trees and might benefit from adaptation.

The diurnal measurements confirm that around midday watertress is most obvious, showing largest differences betweentressed and non-stressed trees. Under severe stress levels diurnalatterns of transpiration, stomatal conductance, and photosyn-hesis related measurements shift, with highest values occurringuring the early hours of the day.

Because of their dependence on climate, the absolute values ofhe physiological measures are of little value for monitoring watertatus except at levels of severe stress (Fernandez and Cuevas,010; Jones, 2007). To be more useful, the data must be comparedo measurements from non-stressed trees or must be somehowormalized to climate. Climate-normalization is common for tran-piration, where ET0 and crop coefficient values are used toompute potential transpiration. This is also the case for CWSIhere canopy temperature is normalized to a climate-determined

heoretical range of leaf temperatures. A similar normalizationould be useful for parameters such as stomatal conductance and

tem water potential.An alternative to compensate for the influence of climate is by

ontinuous monitoring, where measurements can be compared tohemselves and trends analyzed over time to benefit water man-gement decision making. This is basically what is done in theurgor-related PPA measurements, where increases compared tonormal’ indicate onset of stress. Additional plant-based monitoring

ethods with potential for continuous measurements and similarata analysis and utilization are sap flow (Fernández et al., 2008;reven et al., 2009) and trunk or stem size (Moriana and Fereres,

002; Fernandez and Cuevas, 2010). Higher reliability may pos-ibly be achieved by combining methods (Fernandez and Cuevas,010).

Under field conditions, the soil volume available for individualrees is likely to be larger than that in the lysimeters of this exper-

anagement 98 (2010) 124–133

iment. A larger buffer for root available water would likely reducethe intensity of response to drying conditions and possibly give thetree time to apply physiological drought response mechanisms. Theresults from this study were from not yet fruiting young olive treeswith very high vegetative growth rates. Use of the tested methodsfor stress avoidance management is therefore applicable in youngnon-bearing orchards and in orchards in non-reproductive pheno-logical stages where high vegetative growth rates are desirable.The findings need to be confirmed and calibrated with fruit bear-ing trees as well. Several differences are expected for bearing trees.Firstly, that both water consumption and vegetative growth willvary as a function of crop load (Dag et al., 2010). And secondly,since water status will affect quality as well as yield (Dag et al.,2008), that some level of water stress at certain phenological stageswill be desirable to optimize oil production. The methods are there-fore expected to become relevant for mature olive orchards to aidin maintaining target stress levels. These will, however, first needto be determined quantitatively and cumulatively on the whole-tree-scale, in order that they can be translated into physiologicalwater stress management targets defining water stress durationand severity.

5. Conclusions

Drought lead to complete utilization of plant-available waterin soil of lysimeters and to substantial decreases in physiologicalprocesses, but not to their complete shut down. In young olives, aone-time drought event with near total soil water depletion lastingonly 6 days lead to permanent significant reductions in whole-tree evapotranspiration rates and apparent tree size. All testedmethods for direct and indirect monitoring responses to droughtwere successful in recording entry into and exit out of short-term severe water stress. Leaf and stem-based hand measuredstem water potential, stomatal conductance, and fluorescence-based photosynthesis estimations were particularly well correlatedwith whole-tree actual ET. These results represent a first step indetermining thresholds for desired water status in olive trees,measurable by physiological, small-scale monitoring, includingconsiderations of whole-tree and long-term water stress effects.To complete this, continued work on trees in fruit bearing stages isnecessary. Ease and/or expense of each monitoring method obvi-ously influence its potential for ultimate adoption as a managementtool. For this reason, and in spite of remaining technologi-cal and data interpretation challenges, remotely sensed canopytemperature-based monitoring and methods that can be made con-tinuously, like the leaf patch clamp pressure probe, seem promisingand deserve further development and investigation.

Acknowledgement

The building and maintaining of the lysimeter facility was sup-ported generously by the JCA Charitable Foundation (ICA Israel).

References

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration: guide-lines for computing crop water requirements Irrig. Drain. Paper 56. UN-FAO,Rome.

Ben-Gal, A., Shani, U., 2002. A highly conductive drainage extension to control thelower boundary condition of lysimeters. Plant Soil 239, 9–17.

Ben-Gal, A., Agam, N., Alchanatis, V., Cohen, Y., Yermiyahu, U., Zipori, I., Presnov, E.,Sprintsin, M., Dag, A., 2009. Evaluating water stress in irrigated olives: correla-tion of soil water status, tree water status, and thermal imagery. Irrig. Sci. 27,

367–376.

Berenguer, M.J., Vossen, P.M., Grattan, S.R., Connell, J.H., Polito, V.S., 2006. Tree irriga-tion levels for optimum chemical and sensory properties of olive oil. HortScience41, 427–432.

Bergstrom, L., 1990. Use of lysimeters to estimate leaching of pesticides in agricul-tural soils. Environ. Pollut. 67, 325–347.

ater M

C

D

D

D

FF

F

F

F

G

G

G

H

H

I

J

J

A. Ben-Gal et al. / Agricultural W

orwin, D.L., 2000. Evaluation of a simple lysimeter-design modification to minimizesidewall flow. J. Contam. Hydrol. 42, 35–49.

ag, A., Ben-Gal, A., Yermiyahu, U., Basheer, L., Nir, Y., Kerem, Z., 2008. The effectof irrigation level and harvest mechanization on virgin olive oil quality in atraditional rain-fed ‘Souri’ olive orchard converted to irrigation. J. Sci. Food Agr.88, 1524–1528.

ag, A., Bustan, A., Avni, A., Tzipori, I., Lavee, S., Riov, J., 2010. Timing of fruit removalaffects concurrent vegetative growth and subsequent return bloom and yield inolive (Olea europaea L.). Sci. Hortic. 123, 469–472.

e Wit C.T., 1958. Transpiration and crop yield. Versl. Landbouwk. Onderz. 64.6 Inst.of Biol. and Chem. Res. on Field Crops and Herbage, Wageningen, The Nether-lands.

ernández, J.E., Moreno, F., 1999. Water use by the olive tree. J. Crop Prod 2, 101–162.ernandez, J.E., Diaz-Espejo, A., Infante, J.M., Duran, P., Palomo, M.J., Chamorro, V.,

Giron, I.F., Villagarcia, L., 2006. Water relations and gas exchange in olive treesunder regulated deficit irrigation and partial rootzone drying. Plant Soil 284,273–291.

ernández, J.E., Green, S.R., Caspari, H.W., Diaz-Espejo, A., Cuevas, M.V., 2008. Theuse of sap flow measurements for scheduling irrigation in olive, apple and Asianpear trees and in grapevines. Plant Soil 305, 91–104.

ernandez, J.E., Cuevas, M.V., 2010. Irrigation scheduling from stem diameter varia-tions: A review. Agr. Forest Meteorol. 150, 135–151.

lury, M., Yates, M.V., Jury, W.A., 1999. Numerical analysis of the effect of the lowerboundary condition on solute transport in lysimeters. Soil Sci. Soc. Am. J. 63,1493–1499.

iorio, P., Sorrentino, G., d’Andria, R., 1999. Stomatal behaviour, leaf water status andphotosynthetic response in field-grown olive trees under water deficit. Environ.Exp. Bot. 42, 95–104.

reven, M., Neal, S., Green, S., Dichio, B., Clothier, B., 2009. The effect of drought onwater use, fruit development and oil yield from young olive trees. Agr. WaterManage. 96, 1525–1531.

uerfel, M., Baccouri, O., Boujnah, D., Chaïbi, W., Zarrouk, M., 2009. Impacts of waterstress on gas exchange, water relations, chlorophyll content and leaf structurein the two main Tunisian olive (Olea europaea L.) cultivars. Sci. Hortic. 119,257–263.

illel, D., Gairon, S., Falkenflug, V., Rawitz, E., 1969. New Design of a low-costhydraulic lysimeter system for field measurement of evapotranspiration. Isr.J. Agric. Res. 19, 57–63.

siao, T., 1973. Plant responses to water stress. Ann. Rev. Plant Physiol. 24, 519–570.

niesta, F., Testi, L., Orgaz, F., Villalobos, F.J., 2009. The effects of regulated and con-

tinuous deficit irrigation on the water use, growth and yield of olive trees. Eur.J. Agron. 30, 258–265.

ones, H.G., 1992. Plants and Microclimate: A Quantitative Approach to Environmen-tal Plant Physiology. Cambridge University Press, Cambridge.

ones, H.G., 2004. Irrigation scheduling: advantages and pitfalls of plant-based meth-ods. J. Exp. Bot. 55, 2427–2436.

anagement 98 (2010) 124–133 133

Jones, H.G., 2007. Monitoring plant and soil water status: established and novelmethods revisited and their relevance to studies of drought tolerance. J. Exp.Bot. 58, 119–130.

Marek, T., Piccinni, G., Schneider, A., Howell, T., Jett, M., Dusek, D., 2006. Weighinglysimeters for the determination of crop water requirements and crop coeffi-cients. Appl. Eng. Agric. 22, 851–856.

Maxwell, K., Johnson, G.N., 2000. Chlorophyll fluorescence–a practical guide. J. Exp.Bot. 51, 659–668.

Moriana, A., Fereres, E., 2002. Plant indicators for scheduling irrigation of young olivetrees. Irr. Sci. 21, 83–90.

Naor, A., 2006. Irrigation scheduling and evaluation of tree water status in deciduousorchards. Hort. Rev. 32, 111–166.

Pérez-López, D., Gijón, M.C., Moriana, A., 2008. Influence of irrigation rate on therehydration of olive tree plantlets. Agr. Water Manage. 95, 1161–1166.

Ramos, A.F., Santos, F.L., 2009. Water use, transpiration, and crop coefficients forolives (cv. Cordovil), grown in orchards in Southern Portugal. Biosyst. Eng. 102,321–333.

Sepulcre-Cantó, G., Zarco-Tejada, P.J., Jiménez-Munoz, J.C., Sobrino, J.A., de Miguel, E.,Villalobos, F.J., 2006. Detection of water stress in an olive orchard with thermalremote sensing imagery. Agric. Forest Meteorol. 136, 31–44.

Shackel, K.A., Ahmadi, H., Biasis, W., Buchner, R., Goldhamer, D., Gurusinghe, S.,Hasey, J., Kester, D., Krueger, B., Lampininen, B., McGourty, G., Micke, W.,Mitcham, E., Olson, B., Pelletrua, K., Philips, H., Ramos, D., Schwankl, L., Sibbett,S., Snyder, R., Southwick, S., Stevenson, M., Thorpe, M., Weinbaum, S., Yeager, J.,1997. Plant water status as an index of irrigation need in deciduous fruit trees.HortTechnology 7, 23–29.

Sofo, A., Manfreda, S., Fiorentino, M., Dichio, B., Xiloyannis, C., 2008. The olive tree: aparadigm for drought tolerance in Mediterranean climates. Hydrol. Earth Syst.Sci. 12, 293–301.

Tognetti, R., d’Andria, R., Morelli, G., Alvino, A., 2005. The effect of deficit irrigation onseasonal variations of plant water use in Olea europaea L. Plant Soil 273, 139–155.

Tognetti, R., d’Andria, R., Acchi, R., Lavini, A., Morelli, G., Alvino, A., 2007. Deficit irri-gation affects seasonal changes in leaf physiology and oil quality of Olea europaea(cultivars Frantoio and Leccino). Ann. Appl. Biol. 150, 169–186.

Tripler, E., Ben-Gal, A., Shani, U., 2007. Consequence of salinity and excess boron ongrowth, evapotranspiration and ion uptake in date palm (Phoenix dactylifera L.,cv. Medjool). Plant Soil 297, 147–155.

Van Bavel, C.H.M., 1961. Lysimeter measurements of evapotranspiration rates in theeastern United States. Soil Sci. Soc. Am. Pro. 25, 138–141.

Westhoff, M., Reuss, R., Zimmermann, D., Netzer, Y., Gessner, A., Geßner, P., Zim-mermann, G., Wegner, L.H., Bamberg, E., Schwartz, A., Zimmermann, U., 2009.

A non-invasive probe for online-monitoring of turgor pressure changes underfield conditions. Plant Biol. 11, 701–712.

Zimmermann, D., Reuss, R., Westhoff, M., Geßner, P., Bauer, W., Bamberg, E., Ben-trup, F.-W., Zimmermann, U., 2008. A novel, non-invasive, online-monitoring,versatile and easy plant-based probe for measuring leaf water status. J. Exp. Bot.59, 3157–3167.